Hacker News Reader

By smuggling and bias I meant in an ML model. Smuggling was a bit informal, but referring to models overfit on unintended features or artifacts.

View on HN

but we have well established ways to deal with those... test/validation sets, n-fold validation, etc.

Even if there was some overfitting or data contamination that was undetected, the result would most probably still be better than a hand-made decision tree over the same data...

Hand-made decision trees are open to inspection, comprehension, and adaption. There is no way to adapt an opaque ML model to new findings / an experimental treatment except by producing a new model.

Ok, until you can sue the AI you need to find a doctor ok putting their license behind saying “I have no idea how this shiny thing works”. There are indeed some that will, but not a consensus.

Great. Lets have the few that are willing to put their license behind that do so, and then a study showing that those people get better cure rates than those who do not...

And then the decision tree can be rewritten as "Do ${New_method}".

Reply on news.ycombinator.com